There are several programs that can help in these calculations, but the key of these, is that they allow entering the properties of the explosives to be used and the geomechanical characteristics of the rock.
If you want to go through some math, there are several models to "predict" fragmentation; most of them use the Rosin-Rammler function and variations of the probability functions with additional parameters to provide a degree of control over the fines generation. The first published one is Kuz-Ram (C. Cunningham, FragBlast 1, Lulea 1983 and FragBlast 2, Keystone 1987); JKMRC's model presented an improvement in the 90's, assessing fines in a better way. Of the these models, the currently most accepted is Swebrec (Finn Ouchterlony), which is a particularly well described function that can be used to generate a size distribution from a few parameters, and works for fragmentation by blasting and crushing. (I'm attaching one of many Finn's paper on this model).